Structure of subgraphs results

Nick Strayer

6/18/2020

Finding all subgraphs

First we will run subgraph finding algorithm across the associations

subgraphs <- association_pairs %>% head(10000) %>% calculate_subgraph_structure()

subgraphs %>%
  select(-subgraphs) %>%
  head() %>%
  knitr::kable()

step	n_edges	strength	n_nodes_seen	n_subgraphs	max_size	rel_max_size	avg_size	avg_density	n_triples
1	1	0.0916161	2	1	2	1.0000000	2.0	1.0000000	0
2	2	0.0836374	3	1	3	1.0000000	3.0	0.6666667	1
3	3	0.0749584	5	2	3	0.6000000	2.5	0.8333333	1
4	4	0.0729044	6	2	3	0.5000000	3.0	0.6666667	2
5	5	0.0728354	6	2	3	0.5000000	3.0	0.8333333	2
6	6	0.0704478	7	2	4	0.5714286	3.5	0.6666667	2

Visualizing subgraphs search

Next we can investigate the subgraph structure over search

min_rel <- subgraphs %>%
  filter(rel_max_size == min(rel_max_size)) %>%
  tail(1)

max_num_subgraphs <- subgraphs %>%
  filter(n_subgraphs == max(n_subgraphs)) %>%
  tail(1)

max_num_triples <- subgraphs %>%
  filter(n_triples == max(n_triples)) %>%
  tail(1)

subgraphs %>%
  # filter(rel_max_size < 0.5) %>%
  select(
    strength,
    n_subgraphs,
    max_size,
    rel_max_size,
    avg_density,
    n_triples,
    step
  ) %>%
  pivot_longer(-step) %>%
  ggplot(aes(x = step, y = value)) +
  geom_step() +
  geom_vline(xintercept = min_rel$step, color = 'orangered') +
  geom_vline(xintercept = max_num_subgraphs$step, color = 'forestgreen') +
  geom_vline(xintercept = max_num_triples$step, color = 'steelblue') +
  facet_grid(rows = vars(name), scales = "free_y")

node_info <- virus_host_viruses %>%
  rename(id = virus_id) %>%
  mutate(color = ifelse(type == "RNA", "orangered", "steelblue"))

visualize_subgraph_structure(
  association_pairs,
  node_info = node_info,
  subgraph_results = subgraphs,
  trim_subgraph_results = TRUE
)